Functional magnetic resonance imaging (FMRI) is a new, non-invasive imaging technique which locates and displays regional physical changes associated with brain activity. Because of the safety, availability and comparatively low cost of MRI scanning, FMRI has the potential to make functional brain mapping a widely available procedure for diagnosis and management of brain illness. The immediate goal of this project is to develop FMRI as a technique for investigating the cerebral organization of language. Achievement of this goal requires that a variety of standardized activation procedures be developed and that precise descriptions be obtained of the salient variables influencing FMRI responses. A systematic series of activation experiments is proposed, based on techniques drawn from psychophysics, cognitive psychology, and neuropsychology. It is hypothesized that elementary acoustic stimulus variables including rate, intensity, and duration affect the magnitude and spatial distribution of FMRI responses in auditory sensory regions of the cortex. The first specific aim of this project is to measure the relationships between FMRI responses and these stimulus variables. An understanding of these stimulus effects will enable the design and interpretation of subsequent groups of experiments using more complex tasks. It is hypothesized that the auditory cortical regions concerned with speech perception possess an internal functional organization for processing linguistic and non-linguistic features of speech. Consequently, a second specific aim of the project is to map these regions in an effort to clarify the anatomic basis of early speech processing. These studies will exploit the phenomenon of selective attentional modulation to identify functionally specialize areas for low- level acoustic processing, phonetic perception, intonation, and voice discrimination. Finally, it is hypothesized that activation in higher association areas depends on cognitive processing demands distinct from the physical stimulus. The final specific aim of this project is to characterize the FMRI responses associated with cognitive language behaviors which require phonologic and semantic processing functions. Component systems mediating phonetic discrimination, phonological manipulation, and verbal working memory will be mapped using a set of contrasting subtraction paradigms. The structure of semantic representation in posterior association areas will be investigated by manipulating category and modal attributes of retrieved information during a semantic decision task. The role of the left frontal lobe in semantic and phonological functions will be investigated using a hierarchical set of task subtraction experiments. In achieving these aims, this project will lay the foundation for a broader, long-term goal to use FMRI in improving the diagnosis and treatment of brain diseases affecting language function, including dementia, stroke, dyslexia, and epilepsy.